Newbie: in-addr.arpa file for a C Class

Kevin Darcy kcd at daimlerchrysler.com
Mon Jun 7 22:40:18 UTC 2004


belacyrf wrote:

>Can anyone point me in the right direction.  We have a couple reverse
>files 3.2.1.rev for a few of our c classes.  In these files we specify
>the individual reverse entries for each IP.
>
>How do I set up a single in-addr.arpa entry for a full C Class.
>
>For example.  I just want to set up an "in-addr.arpa" file for the
>3.2.1 c class... I dont want to do it by IP.  If I'm not making sense,
>I'm sorry, I'm pretty new. thanks
>
Do you want every address in that /24 range (is it really a "Class C", 
or just a /24? -- sorry, this is a pet peeve of mine) to resolve to the 
same PTR record? You could do that with a wildcard PTR record, i.e.

* in ptr foo.example.com.

But why would you do this? It basically makes reverse lookups for that 
/24 meaningless. You may think you need this in order to humor reverse 
lookups for addresses in that range, but experience tells me you can 
probably achieve this same "humoring" by having an empty reverse zone 
file, where each of those reverse lookups returns "no such domain name". 
This would be less confusing than a wildcard.

You may perhaps not even need the reverse zone at all. If the problem 
you're trying to solve is that you have a private address range (e.g. 
192.168.*.*) and queries for addresses in this particular subrange seem 
to "hang" and eventually time out, then it may be because you are 
forwarding queries by default to the Internet: reverse lookups for 
private addresses don't fare very well on the Internet (you shouldn't be 
polluting the Internet DNS infrastructure with those queries either). 
You should be able to turn off this forwarding selectively by putting 
"forwarders { };" in the zone definition for the highest-level reverse 
zone of your private address space (e.g. 168.192.in-addr.arpa).

                                                                         
                                          - Kevin




More information about the bind-users mailing list